The Use of Financial Spreads as Indicators of Real Activity

A number of recent papers in the US have suggested that financial spreads are useful indicators of real activity. Stock and Watson (1989) is perhaps the most cited, but work with similar general findings is reported by Friedman and Kuttner (1991), and Bernanke (1990) among others. These papers report tests for information (by which read statistical significance) in financial spreads in a multi-variate dynamic model of output. The models proposed for output and methods of estimating these differ between the different papers, and these are described more fully below. But in what has been a variety of approaches each undertaken within a VAR framework, persuasive evidence has appeared that financial spreads may have an informational role. Of course there are questions about the interpretation to be put on these empirical findings; both in the interpretation they may have for economic behaviour and, relatedly, for the policy implications they may have. Answers to both depend in part on the spreads themselves, as discussed more fully below. The US exercises cited above have focused on the spread between yields on commercial paper and treasury bills. Other spreads figuring in the empirical work have included the long corporate bond-government bond yield differential (Bernanke, 1983; Davis, 1992); the yield curve differential (Estrella and Hardouvelis, 1989; Laurent, 1988, 1989; Bernanke and Blinder, 1992; Mishkin, 1989; Browne and Manasse, 1989) and, in the UK only, reverse yield gaps (bond less equity yield) (see Davis and Henry, 1992a, and Davies and Shah, 1992).